Available online at www.notulaebotanicae.

ro
Print ISSN 0255-965X; Electronic 1842-4309
Not Bot Horti Agrobo, 2011, 39(2):42-48 Notulae Botanicae Horti Agrobotanici
Cluj-Napoca

Population Density and Spatial Distribution Pattern of Hypera

postica (Coleoptera: Curculionidae) in Ardabil, Iran
Mona MORADI-VAJARGAH1 , Ali GOLIZADEH1* , Hooshang RAFIEE-
DASTJERDI1 Myron P. ZALUCKI2 , Mehdi HASSANPOUR 1 , Bahram NASERI1
1
University of Mohaghegh Ardabili, Faculty of Agriculture, Department of Plant Protection, Daneshgah
Street, P.O.Box 179, Ardabil, Iran; golizadeh@uma.ac.ir (*corresponding author)
2
The University of Queensland, School of Biological Sciences, St. Lucia, Queensland 4072, Australia

Abstract

The alfalfa weevil, Hypera postica (Gyllenhal), feeds almost exclusively on alfalfa, Medicago sativa L. in most region of the world where
forage crop is grown. It has been investigated the population density and spatial distribution of alfalfa weevil on alfalfa in Ardabil during
2010. Using a 0.25 m2 quadrate sample unit a reliable sample size was 65, with maximum relative variation of 15%. The relative variation
(RV) of the primary sampling data was 13.6. The highest population density of the alfalfa weevil was recorded on 17th April. To estimate
the spatial distribution pattern of this pest, data were analyzed through index of dispersion, Lloyds mean crowding, Morisitas index
and two regression models (Taylors Power Law and Iwaos Patchiness Regression). Taylors model showed an aggregated distribution
pattern for all life stages. Iwaos patchiness regression indicated that larvae, adult and total life cycle had aggregated spatial distribution
(tc < tt), while pupae of alfalfa weevil exhibited a random pattern. The index of dispersion and Lloyds mean crowding methods indicated
an aggregated distribution for this insect. Spatial distribution parameters of this species are used to outline a sampling program as well
as to estimate population density of H. postica development stages. Optimum sample sizes for estimates of larval density, at three levels
of precision, are presented.
Keywords: alfalfa weevil, distribution pattern, Iwaos patchiness regression, population fluctuation, sampling program, Taylors power
law

Introduction port that 60% of alfalfa damage belongs to alfalfa weevil

Alfalfa or lucerne, Medicago sativa L., is one of the
important forage crops cultivated in most regions of the
world. It is an extremely adaptable plant and can be grown
under a broad range of environmental conditions. The crop
is attacked by a diversity of insect pests with the alfalfa wee-
vil, Hypera postica (Gyllenhal), the most destructive pest of
alfalfa in the world (Metcalf and Luckman, 1994). Hypera
postica is a typical oligophagous insect that feeds almost
exclusively on leguminous plants of the genus Medicago,
although it occasionally may feed on a few species of relat-
ed genera including: Melilotus, Trifolium, and Trigonella
in most region of the world where alfalfa is grown. Both
larvae and adults of alfalfa weevil are voracious feeders
damaging terminals, foliage and new crown shoots. Fields
heavily infested may appear silver or white, as most leaves
are skeletonized or consumed entirely. Severe weevil pres-
sure can totally destroy plants (Fick, 1976). In Iran, Afshar
(1937) reported alfalfa weevil, known as Phytonomus vari-
abilis (Herbest), for the first time. In addition he described
its morphological characteristics, distribution and agricul-
tural importance (Karimpour, 1994). Injury levels caused
by H. postica are not exactly clear in Iran but there is a re-
(Khanjani and Pourmirza, 2004). According to studies in
Iran, among alfalfa pests, H. postica is the most important,
as all its life stages feed on alfalfa leaves and stems, but the
main damage is caused by 3rd and 4th instars. If there are
more than 50 larvae of different instars/m2 in a field, all
leaves of alfalfa will be destroyed and the whole of field
will have a white appearance. Sometimes feeding activity
of alfalfa weevil leads to total destruction of alfalfa crop
especially in its first harvest (Karimpour, 1994).
Estimating population densities of pest arthropods is
the most important part of basic research in agricultural
ecosystems and one of the main components in pest man-
agement programs (Kogan and Herzog, 1980); many oth-
er properties are derived from the density measure (Pedigo
and Buntin, 1994). A reliable sampling program to esti-
mate density should include a proper sampling time, sam-
pling unit and sampling size in which the determination
of spatial distribution is crucial (Boeve and Weiss, 1998;
Pedigo and Buntin, 1994; Southwood and Henderson,
2000). Sampling programs can be used in assessing crop
loss (Hughes, 1996), studying the population dynamics of
pests ( Jarosik et al., 2003; Wilson, 1994) and determina-
tion of levels of pest infestation in order to apply control
 Moradi-Vajargah M. et al. / Not Bot Horti Agrobo, 2011, 39(2):42-48
measures (Arnaldo and Torres, 2005). Although the ob-
jectives of population sampling could differ, the develop-
ment of a sampling procedure requires knowledge of the
spatial distribution of the populations (Binns et al., 2000;
Liu et al., 2002; Southwood, 1978). Knowledge of the
spatial distribution of insects is important in understand-
ing the biology and ecology of a species and the basis for
the development of sampling protocols (Bins and Nyrop,
2000).
Several methods are employed in the sampling of alfal-
fa weevil species in crops, some of them being the washing
method (Kuhar et al., 2003), sweep-net sampling (South-
wood, 1978), and the plastic-bag shake method (Higgins et
al., 1991). The latter is the most commonly recommended
(Kuhar et al., 2003).
Methods that are commonly used to describe disper-
sion of arthropods populations have been summarized by
Southwood (1978). Several estimates based on the disper-
sion parameter k of the negative binomial distribution and
on the relationship between variance and mean are used as
indices of aggregation (Davis, 1994; Krebs, 1999; Ludwig
and Reynolds, 1988; Southwood, 1978). Sampling plans
based on these indicators can minimize variation of sam-
pling precision (Kuno, 1991). Having information about
density and changes in population of H. postica during the
season, identification of factors affecting population fluc-
tuations and determination of their effects will help alfalfa
producers in management of this pest. Knowing the spa-
tial distribution of an insect pest is central to design of a
management program (Cho et al., 2001; Wearing, 1988).
Many of these aspects of pests ecology may be region spe-
cific, so it is necessary to do this study in Ardabil region
where there is no reported similar research. Hence, popu-
lation changes of alfalfa weevil and its distribution pattern
were determined in the current study.
Materials and methods
The present study was conducted in an experimental
field of Mohaghegh Ardabili University in the suburb of
Ardabil, Iran from April to August 2010. A field of 5000
m2 was selected for sampling. Irrigation was performed
on a weekly basis. Plants were grown using standard ag-
ronomic practices recommended for Ardabil region. No
insecticides were applied to the field for alfalfa weevil con-
trol in the spring of 2010.
Development of sampling plans
A sampling quadrate was selected as a sample unit as
developed by Hilburn (1985). The density of different bi-
ological stages of alfalfa weevil was estimated by sampling
twice a week from March to August until first alfalfa har-
vest. On average 65 samples of plant matter were collected
from the field using a 0.25 m2 sampling quadrate and were
shaken into a bag to expose alfalfa weevil larvae, pupae and
adults for counting (Kuhar et al., 2003), so the number of
43
alfalfa weevils life stages were recorded. A sub-sample of
20 shaken stems was placed in a modified Berlese funnel to
estimate the proportion of larvae remaining on stems af-
ter the plastic bag shake method for each sample (Higgins
et al., 1991). This proportion was added to the total lar-
val density of each sample. The fundamental principle of
random sampling was observed so that all locations could
have an equal chance of being sampled. This was achieved
by moving in field diameters.
Relative variation (RV) was employed to assess the ef-
fectiveness of the sampling method. RV for the sampling
data was calculated as follows: RV=(SE/m)100 , where SE
is standard error of mean and m is the mean of primary
sampling data. The reliable sample size was determined
using the following equation: N=(ts/dm)2, where, N =
sample size, t = t-student, s = standard deviation, d = de-
sired fixed proportion of the mean and m = the mean of
primary data (Pedigo and Buntin, 1994).
Spatial distribution pattern
The spatial distribution of H. postica among the 65 sam-
ple units was determined by five commonly used methods:
index of dispersion, Taylors power law, Iwaos patchiness
regression, Morisitas coefficient of dispersion and finally
Lloyds mean crowding.
Index of dispersion
Dispersion of a population can be classified through a
calculation of the variance to mean ratio; namely: s2/m=1
random, < 1 regular and > 1 aggregated. Departure from
a random distribution can be tested by calculating the
index of dispersion (ID), where n denotes the number of
samples:
ID = (n-1)s2/m
ID is approximately distributed as x2 with n-1 degrees
of freedom. Values of ID which fall outside a confidence in-
terval bounded with n-1 degrees of freedom and selected
probability levels of 0.95 and 0.05, for instance, would in-
dicate a significant departure from a random distribution.
This index can be tested by Z value as follows:
Z = 2I D - (2 - 1) , = n - 1
If 1.96 Z -1.96, the spatial distribution would be
random but if Z < - 1.96 or Z > 1.96, it would be uniform
and aggregated, respectively (Patil and Stiteler, 1974).
Lloyds mean crowding (x*)
Mean crowding (x*) was proposed by Lloyd to indicate
the possible effect of mutual interference or competition
among individuals. Theoretically mean crowding is the
mean number of other individuals per individual in the
same quadrate: x*=m+s2/m-1
As an index, mean crowding is highly dependent upon
both the degree of clumping and population density. To
remove the effect of changes in density, Lloyd introduced
the index of patchiness, expressed as the ratio of mean
 Moradi-Vajargah M. et al. / Not Bot Horti Agrobo, 2011, 39(2):42-48
44
crowding to the mean. As with the variance-to-mean ratio,
the index of patchiness is dependent upon quadrate size
x*/m=1 random, <1 regular and >1 aggregated (Lloyd,
1967).
Morisitas coefficient of dispersion (I)
Reasoning that the diversity of numbers of individu-
als per quadrate could be used as a measure of spatial pat-
tern, Morisita (1962) developed the index I. Suppose n
quadrates are sampled and xi represents the number of
individuals in the ith quadrate. Define as the probabil-
ity that individuals of a randomly drawn pair will come
from the same quadrate. If the xi came from a random or
poisson distribution, the expected value of is 1/n. Then
the index I can be defined as the ratio of to its expected
value assuming a random distribution. Values for I may
range from 0 to n and N is the total number of individuals
sampled in n quadrate. I is calculated as follows and used
to classify population:
x (x - 1)
I =n i i
N(N - 1)
To determine if the sampled population significantly
differs from random, the following large sample test of
Test b = 1: t=(b-1)/SEb and Test = 1: t=(-1)/SE,
where SEb and SE are the standard errors of the slope for
the mean crowding regression. Calculated values are com-
pared with tabulated t-values with n-2 degrees of freedom.
If the calculated t (tc) < t-table (tt), the null hypothesis (b
= 1) would be accepted and spatial distribution would be
random. If tc> tt, the null hypothesis would be rejected and
if b > 1 and b < 1, the spatial distribution would be aggre-
gated and uniform, respectively.
Optimum number of sample units (sample size)
The optimum sample size is the smallest number of
sample units that would satisfy the objectives of the sam-
pling program and achieve the desired precision of esti-
mates. Finding out the Taylor power law and Iwaos regres-
sion coefficients eliminates experimental needs for large
sample size (Ifoulis and Savopoulou-Soultani, 2006). The
optimum number of sample units was derived from for-
mula
2
t
N opt D / 2 am b  2
D
using Taylors power law coefficients and from
2
t D 1
significance can be used: Z=(I-1)/(2/nm2)(1/2), where m
= mean population density per leaf in each sampling date
and n = the number of sample units. By comparing the
value of Z with tabulated values for a random distribution
and reject the hypothesis that the sampled population is
dispersed |Z|>z(/2)randomly if (Pedigo and Buntin,
1994).
Regression methods
For many insect and animal species, Taylor (1961)
found that a power law function could be used to model
the relationship between mean and variance as: s2=amb ,
where s2 is the variance; m the sample mean; a is a scaling
factor related to sample size and b measuring the species
aggregation. When b = 1, < 1 and > 1, the distribution
is random, regular and aggregated, respectively. Through
use of a log transformation, one can estimate the coeffi-
cients with linear regression as: log(s2)=log(a)+blog(m) ,
where a and b are the parameters of the model, estimated
by linearizing the equation after a log-log transformation
(Taylor, 1961).
Iwaos patchiness regression method was used to
quantify the relationship between mean crowding index
(m*) and mean density (m) using the following equation:
m*=+m , where indicates the tendency to crowding
(positive) or repulsion (negative) and reflects the dis-
tribution of population in space and is interpreted in the
same manner as b of Taylors power law (Iwao and Kuno,
1968). Student t-test can be used to determine whether
the colony is composed of single individuals and if colo-
nies are dispersed randomly.
N opt D / 2  E  1
D m
using Iwaos regression method coefficients (Buntin,
1994; Young and Young, 1998), where D represents the
range of accuracy, a and b is Taylor regression coefficients,
and is Iwaos equation coefficients and m is mean den-
sity of populations and t /2 is t-student. It has been cal-
culated the optimum number of sample units with 15, 20,
and 30% confidence interval levels and plotted them using
excel program.
Results and discussion
Sampling and population fluctuation
The results from primary sampling showed that the re-
liable sample size with maximum variation of 15% was 20
samples during the early season and an average of 65 sam-
ples during growing season. The relative variation (RV) of
the primary sampling data was 13.6, which is suitable for
a sampling plan.
Rainfall and climatic conditions in the early growing
season caused the first sampling date to be 12 April. The
population increased from the beginning date and after a
week reached a peak by 17 April at 619 larvae per quad-
rate (Fig. 1). Then the population decreased sharply in the
following week. Thereafter, the population continued to
decrease until the end of the growing season. No H. postica
were observed after 20 July (Fig. 1).
In developing a sampling program for research or man-
agement purposes, it is necessary to determine two char-
acteristic features of any population, its density as well as
its dispersion (Pedigo and Buntin, 1994). Alfalfa weevil
 Moradi-Vajargah M. et al. / Not Bot Horti Agrobo, 2011, 39(2):42-48
population density showed a rising trend during the early
growing season with small larva, first and second instars,
constituting the bulk of the population. The highest pop-
ulation density during the April could be attributed to
unfavorable climatic conditions for natural enemies and
high quality of host plant. As the season progressed the
reduction in host plant quality and higher activities of al-
falfa weevil biocontrol agents could have led to a decreas-
ing population. Moreover, the early high density of larvae
caused a significant reduction of leaf and young stems in
the alfalfa field and this could be another reason for the
following reduction in population density. Along with
warming weather and reducing humidity, population den-
sity decreased reaching near to zero per sampling unit. The
last observation of beetle was made by last week of July.
Considering the changes in H. postica density, it is possible
that timing the first cutting of alfalfa to coincide with peak
alfalfa weevil populations may be a suitable means of con-
trol. Thus, in order to reduce population density, any effec-
tive pest management strategies should be applied during
the early growing season against H. postica.
Fig. 1. Population fluctuation of Hypera postica in alfalfa (mean
SE) during the growing season in 2010 in Arabil, Iran. Each
estimate is based on 65 random 0.25 m2 quadrates
Spatial distribution pattern
The variance to mean ratio were greater than unity
indicating that the spatial distribution of alfalfa weevil
stages in alfalfa field is aggregated (Tab. 1); the calculated
Z-values of each stage was significantly greater than 1.96.
The aggregated distribution suggests that the presence of
an individual at one point leads to an increased probability
of another individual being close.
The Lloyds mean crowding were calculated for H.
postica life stages and x*/m ratio ranged from 1.08 to 2.66
(Tab. 1). Similar conclusions can be made from the results
of Lloyds mean crowding for alfalfa weevil life stages dis-
tribution which indicated that H. postica had an aggre-
gated distribution on alfalfa. The use of dispersion indices
seem to be convenient decision making methods for man-
agement programs because of their easy calculation proce-
dure and simple results (Darbemamieh et al., 2011).
45
Tab. 1. Estimated parameters by variance to mean ratios and
Lloyds mean crowding methods for life stages of Hypera postica
in 2010
Larvae Pupae Adult Total
Variance to mean ratio
S2/m 64.530 3.303 4.286 63.440
ID 1290.611 69.374 42.865 1712.902
Z 44.560 5.376 4.900 51.25
Lloyds mean crowding
x* 88.52 1.74 1.21 103.96
x*/m 1.76 2.66 1.08 1.71
The Morisitas index (I) values for all sampling dates
and for three life stages and total life span of H. postica is
significantly greater than 1 (Tab. 2), indicating the spatial
distribution of this pest was aggregated at most sampling
dates. As Morisitas coefficient estimates spatial distribu-
tion pattern using the mean and variance of each sampling
date separately, so this index is more perfect than disper-
sion index. The detailed knowledge of dispersion in differ-
ent time intervals during growing season would be useful
for research strategies more than management programs
(Darbemamieh et al., 2011). Although, the spatial pattern
of alfalfa weevil in all sampling date was aggregated pat-
tern, changes in distribution pattern during season could
be casued by changes in population density or movement
of larva from the clumped egg locations (Pedigo and Bun-
tin, 1994).
According to Taylors model (Tab. 3), the calculated
intercepts (a) for larvae, pupae, adult and total stages of
H. postica were greater than zero (a = 3.27, 2.58, 3.10 and
3.14 respectively) and all b values from Taylors power law
were significantly >1, indicating an aggregated dispersion
pattern of H. postica stages (Tab. 3). There were significant
relationship between log S2 and log m for all life stages of
H. postica. Moreover, the coefficients of determination for
the Taylors power law ranged from 0.867 to 0.977 (Tab.
4).
Similar to Taylors power law, Iwaos patchiness regres-
sion produced comparable results for the distribution of
alfalfa weevil. Iwaos patchiness regression adequately de-
scribed the relationship between mean crowding (x*) and
mean density (m) for life stages of H. postica (P<0.01, Tab.
4). In this model, values of ranged from 1.116 to 2.370.
The coefficients of determination for the Iwaos patchiness
regression ranged from 0.609 to 0.974. For life stages of
alfalfa weevil, Iwaos intercept (a) was considerably greater
than 0, indicating a very strong tendency to crowding. The
b values of Iwao model were significantly >1 for larvae,
adult and as well as for total life span, indicating that lar-
vae and adult weevils tend to be aggregated. Although the
slope of Iwao model (b) was greater than unity for pupae
stage, it is not significantly different from 1 (tc < tt), indi-
cating that pupae of H. postica are randomly distributed in
the sampling area. This differs from other analysis meth-
ods used in this study.
 Moradi-Vajargah M. et al. / Not Bot Horti Agrobo, 2011, 39(2):42-48

46
Tab. 2. Parameters of Morisitas index and Z calculated for Hypera postica on Alfalfa in 2010 for each sampling date
Larvae Pupae Adult Total
Date
I z I z I z I z
12-Apr 14313095.0 105627.2 - - 15.0 78.3 14333094.3 197625. 7
17-Apr 43273091. 5 43273091.5 - - - - 43273091.5 43273091.5
24Apr 119680805.0 119679622.0 - - - - 119680804.8 119679622.2
29Apr 38506190.0 38506190.0 - - - - 38506190.0 38506190.0
1-May 3538524.4 3538524.4 - - - - 3538524.4 3538524.4
6-May 1475223.9 1475223.8 - - - - 1480501.5 1480501.4
10-May 2164170.8 2164170.8 - -21.1 - - 2174087.1 2174087.0
13-May 1910725.5 1910725.5 298.7 294.5 - - 1958705.9 1958705.8
17-May 526356.4 526356.3 137.1 128.1 - - 531796.7 531796.6
20-May 375706.6 375706.5 77.1 64.1 - - 385859.3 385859.2
24-May 441673.10 441673.0 32.0 19.4 - - 447309.7 447309.6
27-May 231358.5 231358.3 2150.4 2147.9 - - 260763.3 260763.1
30-May 52101. 7 52101.3 180.0 164.2 - - 52666.7 52666.3
2-Jun 102838.6 102838.4 26.7 5.6 0 -63.2 104638.5 104638.2
6-Jun 111451.3 111451.2 705.7 701.2 - - 127215.7 127215.5
9-Jun 54961.0 54960.7 371.4 366.9 - - 60871.5 60871.2
12-Jun 52446.8 52446.5 1044.0 1041.9 - - 64390.0 6439. 7
15-Jun 8323.0 8321.8 33.3 22.8 - - 8889.3 8888.3
19-Jun 338.8 335.1 784.0 779.8 - - 1860.0 1858.0
24-Jun - - 5842.1 5840.4 6399.3 6398.1 24618.1 24617.4
27-Jun 746.7 739.6 7589.2 7588.1 965.7 961.2 10352.2 10351.4
30-Jun - - 1612.8 1610.3 1098.9 1095.6 3948.2 3946.7
4-Jul - - 96.0 83.4 100.0 89.5 363.6 357.9
7-Jul - - 1733.2 1732.7 1286.8 1286.0 54398.1 54397.8
11-Jul - - 315.0 307.1 627.7 622.8 2133.3 2130.3
14-Jul - - 0 -63.2 4343.5 4342.1 4345.5 4344.2
18-Jul - - 20.0 -11.6 137.1 128.1 177.8 170.8
20-Jul - - - - 1418.3 1415.7 1418.3 1415.7
24-Jul - - - - - - - -
28-Jul - - - - - - - -
Since in regression methods the mean and variance of lating the spatial distribution could be different. Based on
each sampling time was used separately, therefore the Tay- Southwood (1978), when a population in an area becomes
lors power law and Iwaos patchiness were more accurate sparse, the chance of an individual occurring in any sample
than the variance to mean ratio method. The two regres- unit is so low that the distribution is in effect random.
sion techniques (Taylors Power Law and Iwaos patchiness Spatial distribution of the alfalfa life stages using differ-
regression) have been widely used to evaluating dispersal, ent analytical methods all indicated an aggregated pattern.
data normalizing for statistical analysis, and developing This behavior has been mentioned for many other insects
sampling protocols for many insects (Davis, 1994; Deli- and mites (Darbemamieh et al., 2011; Geiger and Daane,
georgidis et al., 2002). The varied conclusions for disper- 2001; Hamilton and Hepworth, 2004; Sedaratian et al.,
sion of a life stage pattern from two regression methods 2010). Zahiri et al. (2006) analyzed the spatial distribution
suggest that different statistical methods accuracy in calcu- of alfalfa weevil and some of its natural enemies by Taylos
Tab. 3. Spatial Distribution of Hypera postica on alfalfa during Tab. 4. Spatial Distribution of Hypera postica on alfalfa during
2010 using Taylors power law analysis 2010 using Iwaos patchiness regression analysis
Life stages a b SE b r2adj P reg tc tt Life stages SE r2adj P reg tc tt
Larvae 3.273 1.355 0.089 0.920 0.00 3.988 2.093 Larvae 4.276 1.148 0.035 0.974 0.00 4.228 2.093
Pupae 2.582 1.300 0.090 0.908 0.00 3.333 2.086 Pupae 1.023 1.467 0.252 0.609 0.00 1.853 2.086
Adult 3.097 1.416 0.068 0.977 0.00 6.117 2.262 Adult 0.151 2.370 0.105 0.946 0.00 13.047 2.262
Total 3.140 1.288 0.096 0.867 0.00 3.00 2.056 Total 5.754 1.116 0.036 0.972 0.00 3.222 2.056
 Moradi-Vajargah M. et al. / Not Bot Horti Agrobo, 2011, 39(2):42-48
power law in Hamedan, west Iran and reported clumped
dispersion patterns for life stage of H. postica and random
dispersion for its natural enemies. Muker and Guppy
(1972) considered H. postica to be aggregated on alfalfa
which is the same as the present result. Despite parameters
such as rate of population growth and reproduction that
vary from one generation to another, spatial distribution
is partially constant and is a characteristic of the species
(Taylor, 1984). The behavioral patterns and environment
could be determinant the spatial distribution of popula-
tion individuals in an ecosystem. Different plant varieties
can affect the spatial distribution of an insect (Sedaratian
et al., 2010), therefore, further work is needed to verify the
spatial distribution of this pest in commercial alfalfa fields
and on different varieties of alfalfa. The dispersion pat-
tern of alfalfa weevil identified in this study can be used in
sampling program as well as in proper estimation of their
population densities in protection management of alfalfa.
Optimum number of sample units
The optimal sample size was calculated using Taylors
and Iwaos coefficient, and the relationship between op-
timum number of sample unit and mean number of H.
Optimal sample size
Mean larvae per sample unit
Fig. 2. Relationship between optimal sample size (0.25 m2 quad-
rates) and mean number of Hypera postica larvae, for three levels
of precision, based on Taylors power law (A) and Iwaos regres-
sion method (B) coefficient
47
postica larvae, for 15, 20 and 30% level of precision was
calculated (Fig. 2). Comparison of the two different for-
mulae used for calculating the optimal sample unit size
showed a similarity between the lines calculated using the
both regression methods (Fig. 2). The optimal sample size
suggested by each model is typically higher at low popula-
tion levels. The lower estimate of the sample size was cal-
culated by using Taylors equation, for larval stages. Prin-
cipally, Taylors method reduces the necessary sample size
when compared with Iwaos method (Darbemamieh et al.,
2011; Ifoulis and Savopoulou-Soultani, 2006). Taylors
power law model usually better fits the spatial dispersion
than Iwaos model to accomplish a desired precision of es-
timates (Afshari et al., 2009). These different precision lev-
els (three lines) adopted in this study, could be chosen for
ecological or insect behavioral studies. In order to acquire
a higher level of precision the 15% level could be adopted,
whereas in IPM programs 20% or 30% level would be ac-
ceptable.
References
Afshari A, Soleyman-Negadian E, Shishebor P (2009).
Population density and spatial distribution of Aphis gossypii
Glover (Homoptera: Aphididae) on cotton in Gorgan, Iran. J
Agr Sci Tech 11:27-38.
Arnaldo PS, Torres LM (2005). Spatial distibution and sampling
of Thaumetopoea pityocampa (Lep: Thaumetopoeidae)
populations of Pinus pinaster Ait. in Montesinho, N.
Portugal. Forest Ecol Manage 210:1-7.
Binns MR, Nyrop JP, Werf W (2000). Sampling and monitoring
in crop protection: The theoretical basis for developing
practical decision guides. CABI Publishing, UK.
Boeve PJ, Weiss M (1988). Spatial distribution and sampling
plans with fixed levels of precision for cereal aphids
(Homoptera: Aphididae) infesting spring wheat. Can
Entomol 130:67-77.
Buntin GD (1994). Developing a primary sampling program,
99-11 p. In: Pedigo LP, Buntin GD (Eds.). Handbook of
sampling methods for arthropods in agriculture. CRC, Boca
Raton, FL.
Cho K, Lee JH, Park JJ, Kim JK, Uhm KB (2001). Analysis of
spatial pattern of greenhouse cucumbers using dispersion
index and spatial autocorrelation. Appl Entomol Zool
36(1):25-32.
Darbemamieh M, Fathipour Y, Kamali K (2011). Population
abundance and seasonal activity of Zetzellia pourmirzai
(Acari: Stigmaeidae) and its preys Cenopalpus irani and
Bryobia rubrioculus (Acari: Tetranychidae) in sprayed apple
orchards of Kermanshah, Iran. J Agr Sci Tech 13:143-154.
Davis PM (1994). Statistics for describing populations, 33-54
p. In: Pedigo LP, Buntin GD (Eds.). Handbook of sampling
methods for arthropods in agriculture. CRC Press, Boca
Raton, Florida.
 Moradi-Vajargah M. et al. / Not Bot Horti Agrobo, 2011, 39(2):42-48
48
Deligeorgidis PN, Athanassiou CG, Kavallieratos NG (2002).
Seasonal abundance, spatial distribution and sampling
indices for thrips populations on cotton; a four year survey
from Central Greece. J Appl Entomol 126:343-348.
Fick GW (1976). Alfalfa weevil effects on regrowth of alfalfa.
Agron J 68:809-812.
Geiger CA, Daane KM (2001). Seasonal movement
and distribution of the grape mealybug (Homoptera:
Pseudococcidae): developing a sampling program for San
Joaquin valley vineyards. J Econ Entomol 94:991-301.
Hamilton AJ, Hepworth G (2004). Accounting for cluster
sampling in constructing enumerative sequential sampling
plans. J Econ Entomol 97:1132-1136.
Higgins RA, Rice ME, Blodgett SL, Gibb TJ (1991). Alfalfa
stem-removal methods and their efficiency in predicting
actual numbers of alfalfa weevil larvae (Coleoptera:
Curculionidae). J Econ Entomol 84:650-655.
Hilburn DJ (1985). Population dynamics of overwintering
life stages of the alfalfa weevil, Hypera postica (Gyllenhal).
Virginia Polytechnic Institute and State University, PhD
Diss. Blacksburg.
Hughes GM (1996). Incorporating spatial pattern of harmful
organsim into crop loss models. Crop Prot 15:407-421.
Ifoulis AA, Savopulou-Soultani M (2006). Developing
optimum sample size and multistage sampling plans for
Lobesia botrana (Lepidoptera: Tortricidae) larval infestation
and injury in Northern Greece. J Econ Entomol 99:1890-
1898.
Iwao S, Kuno E (1968). Use of the regression of mean crowding
on mean density for estimating sample size and the
transformation of data for the analysis of variance. Res Popul
Ecol 10:210-214.
Jarosik V, Honek A, Dixon AFG (2003). Natural enemy
ravine revisited importance of sample size for determining
population growth. Ecol Entomol 28:85-91.
Karimpour Y (1994). Studies on the effect of different
insecticides on alfalfa weevil (Hypera postica) and etrimfos
residues in green alfalfa. Tarbiat Moddares University, MSc
Thesis, Tehran.
Khanjani M, Pourmirza AA (2004). A comparison of various
control methods of alfalfa weevil, Hypera postica (Col:
Curculionidae) in Hamadan. J Entomol Soc Iran 1:67-81.
Kogan M, Herzog DC (1980). Sampling methods in soybean
entomology. Springer Verlag, New York.
Krebs CJ (1999). Ecological methodology. 2nd ed. Addison
Wesley Longman Inc, New York.
Kuhar TP, Youngman RR, Laub CA (2003). Alfalfa weevil
(Coleoptera: Curculionidae) population dynamics and
mortality factors in Virginia. Environ Entomol 29(6):1295-
1304.
Kuno E (1991). Sampling and analysis of insect population.
Ann Rev Entomol 36:357-1971.
Liu C, Wang G, Wang W, Zhou S (2002). Spatial pattern of
Tetranychus urticae population in apple tree garden. J Appl
Ecol 13:993-996.
Lloyd M (1967). Mean crowding. J Anim Ecol 36:1-30.
Ludwig JA, Reynolds JF (1988). Statistical ecology-a primer on
methods and computing. John Wiley and Sons Inc, New
York.
Metcalf RL, Luckman WH (1994). Introduction to insect pest
management. 3rd ed. Wiley, New York.
Morisita M (1962). I-index a measure of dispersion of
individuals. Res Popul Ecol 4:1-7.
Muker MK, Guppy JC (1972). Notes on the spatial pattern of
Hypera postica (Coleoptera: Curculionidae) on alfalfa. Can
Entomol 104:1995-1999.
Patil GP, Stiteler WM (1974). Concepts of aggregation and
their quantification: a critical review with some new results
and applications. Res Popul Ecol 15:238-254.
Pedigo LP, Buntin GD (1994). Handbook of sampling methods
for arthropods in agriculture. CRC Press, Florida.
Sedaratian A, Fathipour Y, Talebi AA, Farahani S (2010).
Population density and spatial distribution pattern of
Thrips tabaci (Thysanoptera: Thripidae) on different soybean
varieties. J Agr Sci Tech 12:275-288.
Southwood TRE (1978). Ecological methods with particular
reference to the study of insect populations. 2nd ed. Chapman
and Hall, London.
Southwood TRE, Henderson PA (2000). Ecological methods.
3rd ed. Blackwell Sciences, Oxford.
Taylor LR (1961). Aggregation, variance to the mean. Nature
189: 732-735.
Taylor LR (1984). Assessing and interpreting the spatial
distribution of insect populations. Ann Rev Entomol
29:321-357.
Wearing CH (1988). Evaluating the IPM implementation
process. Ann Rev Entomol 33:17-38.
Wilson LT (1994). Estimating abundance, impact, and
interactions among arthropods in cotton agroecosystems,
475-514 p. In: Pedigo LP, Buntin GD (Eds.). Handbook of
sampling methods for arthropods in agriculture. CRC Press,
Boca Raton.
Young LJ, Young JH (1998). Statistical ecology: a population
perspective. Kluwer Academic Publishers, Norwell, MA.
Zahiri B, Fathipour Y, Khanjani M, Moharramipour S (2006).
Spatial distribution of alfalfa weevil and its natural enemies
in Hamedan. Proceeding of 17th Iranian Plant Protection.
Karaj, Iran, 52 p.